Most firms in the economy charge markups, that is, they sell goods at a price that is larger than the cost of production. Further, larger firms tend to charge higher markups and the overall level of markups in the economy has been rising in recent decades. These facts have become a leading concern for both academics and policy makers. Counter-intuitively, the research on this topic so far has robustly concluded that large firms that charge high markups are too small. The main logic behind this result is that firms restrict the amount they sell in order to be able to charge high prices. It follows that policy makers should subsidize these large firms in order to incentivize them to produce more.

In my research with Gideon Bornstein, we challenge this classic result and show that it is a direct consequence of one key assumption commonly made: firms must offer constant per-unit prices, precluding more sophisticated strategies such as quantity discounts. Using data on consumer-packaged goods, we first show that quantity-dependent pricing is prevalent. Around 90% of sales in this sector are in products that are offered in multiple sizes, where the large size comes with a sizeable price discount. On average, when a package is 10% larger, firms charge 7% less per unit. We then develop a new framework in which firms can offer flexible pricing strategies that more closely resemble the data.

Why would firms price in such a way? In our model, this kind of quantity discount is optimal when firms face different consumers who have different tastes for their product. In order to be able to charge all types of consumers the highest possible price, firms try to differentiate their various size options as much as possible and end up selling too much in the big packages and too little in the small ones. This leads to a new source of inefficiency: different consumers purchasing from the same firm are not being sold the optimal quantity. We find that the new inefficiency due to firms distorting how much they sell to different consumers leads to welfare losses of around 1%. This is more than twice as large as the classic welfare losses across firms that have been studied by previous research.

When looking at total firm sales to all consumers, we show that market shares are actually optimal for all firms. The conclusion that high-markup, large firms are too small and policy makers should subsidize them is no longer valid. On the contrary, such a policy would lead to large welfare losses for consumers. Subsidizing large firms that charge high markups, as would be optimal in an environment in which firms are restricted to charge the same per-unit price to all consumers, would lead to additional welfare losses of 0.4%.

Last, we show that our new approach to modeling firms’ pricing behavior also changes the relationship between markups and the economy’s overall level of labor supply. In the standard setup with constant per-unit prices, the fact that firms charge markups implies that real wages are low, discouraging labor supply by households. This conclusion no longer holds in our model and there is a much smaller role for a general wage subsidy. In follow-up research, we plan to study how firm’s ability to offer quantity-dependent prices interacts with income and spending inequality.